Routine abdominal drainage after distal pancreatectomy: meta-analysis

Department of Surgery, Amsterdam UMC, University of Amsterdam, Cancer Centre Amsterdam, Amsterdam, the Netherlands Department of Surgery, Poliambulanza Hospital Brescia, Brescia, Italy Department of General and Pancreatic Surgery, Pancreas Institute, University of Verona Hospital Trust, Verona, Italy DepartmentofSurgery,RegionalAcademicCancerCentreUtrecht,UniversityMedicalCentreUtrecht, andStAntoniusZiekenhuisNieuwegein,Utrecht, theNetherlands Department of Surgery, Erasmus MC, Rotterdam, the Netherlands

Most studies of drain management combined DP with pancreatoduodenectomy, and are therefore less useful [8][9][10][11][12] .Drain placement may lead to retrograde infection, patient discomfort, or direct damage to blood vessels 13 .A recent multicentre randomized trial 6 demonstrated comparable outcomes with and without routine abdominal drainage after DP.It is unclear, however, whether omitting routine drainage in subgroups with a high risk of POPF would potentially lead to an increased risk of complications.
A systematic review of abdominal drainage after DP specifically is lacking.In this systematic review, the benefits and risks associated with a no-drain strategy versus abdominal drainage after DP were compared.

Methods
A systematic review and meta-analysis was undertaken to compare no drain placement versus routine abdominal drainage in patients undergoing DP.The primary outcome was major morbidity, defined as complications with a Clavien-Dindo grade of III or higher.  1 and Fig. 1).There was no heterogeneity in the primary outcome between the studies.All five studies 6,14-17 reported data on POPF grade B/C.Pooled analysis showed that the POPF rate was lower in the no-drain group compared with the drain group (RR 0.55, 0.42 to 0.72).Readmissions were reported in three studies 1,16,17 , with a lower rate in the no-drain group (RR 0.76, 0.60 to 0.96).
Rates of radiological intervention, postpancreatectomy haemorrhage, delayed gastric emptying, intra-abdominal abscess, surgicalsite infection, reoperation, and 30-day mortality were no different between groups.Detailed results of the meta-analysis are shown in Appendix S4, Table 1, and Fig. 1.

Discussion
No drain placement after DP was associated with a lower rate of major complications (Clavien-Dindo grade at least III), POPF, and readmissions.Rates of radiological intervention and reoperation did not differ.No study has reported on high-risk subgroups.
Five studies were included in the present meta-analysis, which has a high statistical power and effect size by including a large number of patients.In the study by Paulus and colleagues 17 the nodrain group had a lower rate of POPF (0 versus 15 per cent), without differences in other complications.The discrepancy between POPF and other complications in the no-drain group can be explained by use of the older terminology for POPF, which has been updated since then.This why severe morbidity was chosen as primary endpoint in the present study.Mangieri et al. 16 reported a higher rate of POPF grade B/C and readmissions in the drain group.Behrman and co-workers 14 reported no difference between groups in severe morbidity and grade B/C POPF.Correa-Gallego and colleagues also did not find any disadvantages in the no-drain group 15 .The only included randomized multicentre trial, by Van Buren et al., did not find a difference in rate of POPF, but noted comparable rates of radiological intervention between the groups 5 .This trial did not stratify by subgroups such as high-and low-risk POPF.It therefore remains unclear whether the outcomes reported in the present meta-analysis also apply to high-risk subgroups.This meta-analysis has confirmed the findings of Van Buren that a routine drain policy does not protect the patient from additional radiological interventions.
Recently, the first distal fistula risk score was constructed, which includes duct size and pancreatic thickness (M.v. B. E. De Pastena, submitted for publication).This prediction model enables the clinician to determine the risk of POPF, so that appropriate measures can be taken, such as selective drainage in high-risk patients.Future pragmatic multicentre randomized trials including risk-stratified randomization are required before final conclusions can be drawn.
This study had several limitations.Non-randomized studies could have been exposed to selection bias, information bias, and follow-up bias because patients who did not receive drains may have had favourable characteristics leading to omission of drains.The definition of POPF differed between studies.Potential bias was minimized by analysing only POPF grade B/C according to the ISPGS 1 .There was heterogeneity between studies.In most studies, however, no clinically relevant differences were observed between preoperative, perioperative, and postoperative parameters in the two groups, such as stump closure methods.Different stump closure methods could lead to a difference in POPF rate 18 .

Fig. 1
Fig. 1 Meta-analysis of impact of no drain versus drain on outcomes after distal pancreatectomy a Major morbidity, b grade B/C postoperative pancreatic fistula, c radiological intervention, and d readmission.A Mantel-Haenszel fixed-effect model was used for meta-analysis.Risk ratios are shown with 95 per cent confidence intervals.

Table 1
Summary of findings for no drain versus drain